Re: [Gretl-devel] Lee's bkw.gfn and vif (was Re: (now vif-related) Re: new accessor)

Am 13.03.2019 um 23:34 schrieb Riccardo (Jack) Lucchetti: > On Wed, 13 Mar 2019, Sven Schreiber wrote: >> Drawing the connection to the vif-related feature request >> (https://sourceforge.net/p/gretl/feature-requests/101/), does this new >> $result bundle contain everything >> That would be great of course, because it would solve that ticket (and by >> implication, probably would make the bkw.gfn package obsolete). > > Yes, I suppose it does. Well, it's looking good already (tried with the latest snapshot on Windows), but there's one thing that the bkw.gfn package covers and the 'vif' command doesn't, namely nonlinear models. I got an error trying out 'vif' after an 'nls' block. In the doc for 'vif' this limitation doesn't seem to be mentioned. I guess the VIFs themselves do not make much sense outside of ols (or do they?), but these days the vif command also gives you the BKW table, of course. So perhaps could 'vif' be activated for nonlinear models, and give only the BKW table?